CN104601991A - Searching method and device for time-domain predictive parameters - Google Patents
Searching method and device for time-domain predictive parameters Download PDFInfo
- Publication number
- CN104601991A CN104601991A CN201410851138.2A CN201410851138A CN104601991A CN 104601991 A CN104601991 A CN 104601991A CN 201410851138 A CN201410851138 A CN 201410851138A CN 104601991 A CN104601991 A CN 104601991A
- Authority
- CN
- China
- Prior art keywords
- merging patterns
- adama
- best merging
- coding unit
- cost
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The embodiment of the invention provides a searching method and a searching device for time-domain predictive parameters, wherein the searching method for the time-domain predictive parameters comprises the following steps: determining combined-mode optimal movement parameters of a coding unit, and determining an optimal combining mode of the coding unit; determining Adama total cost of the optimal combining mode of the coding unit, wherein the Adama total cost of the optimal combining mode of the coding unit is sum of Adama cost of the optimal combining mode and the index cost of the optimal combining mode, the index cost is product of a lagrange parameter and a bit number corresponding to the optimal combining mode index; while the Adama total cost of the optimal combining mode meets the first ending condition, determining the predictive parameters of the optimal combining mode as optimal time-domain predictive parameters of an existing coding unit. The searching method can be used for reducing complexity of inter-frame coding and improving the coding efficiency.
Description
Technical field
The present invention relates to technical field of video coding, particularly relate to a kind of searching method and device of time domain prediction parameter.
Background technology
HEVC (High Efficiency Video Coding, efficient video coding standard) is video encoding standard of new generation, and its object produced is the demand that solution people increase day by day to vision and acoustical quality.
HEVC has two kinds of interframe encoding modes, one merges (Merge) pattern, the kinematic parameter of the encoded unit that merging patterns utilize space-time adjacent obtains the kinematic parameter (comprising reference key and motion vector) of current coded unit, and Merge index is enrolled code stream; Another kind is motion estimation mode, motion estimation mode is for current coded unit, generally there are 2 reference columns, forward direction reference columns (List0) and backward reference columns (List1), each reference columns can have one or more reference picture, to each in all reference pictures, fill order is to estimation, obtain best forward and backward kinematic parameter, then perform bi-directional motion estimation and obtain best bidirectional-movement parameter, from three, finally select optimum one as optimal movement estimation model.
The searching method of existing a kind of time domain prediction parameter specifically can comprise following process step: the Prediction Parameters first generating maximum 5 Merge patterns, then calculate rate distortion cost or the Adama cost of each Prediction Parameters, therefrom select one of least cost as best Merge pattern, then to each in all reference pictures of current coded unit, fill order is to estimation, obtain best forward and backward kinematic parameter, then perform bi-directional motion estimation and obtain best bidirectional-movement parameter, from best forward direction, optimum a kind of calculation rate distortion cost or Adama cost is selected in backward and bidirectional-movement parameter three, and the cost of the rate distortion of selection result and best Merge pattern or Adama cost are contrasted, determine the Priority Inter-frame coding mode of rate distortion cost or Adama least cost, and using the best time domain prediction parameter of corresponding for this optimum interframe encoding mode time domain prediction parameter as current coded unit.But, computational process all more complicated of motion estimation process and rate distortion cost or Adama cost, therefore, exhaustive all motion estimation processes and to each situation in two kinds of interframe encoding modes carry out rate distortion cost or Adama cost calculating, all need complicated calculating and the amount of calculation of meeting at substantial, because this increasing complexity and the scramble time of coding.
Summary of the invention
Embodiment of the present invention technical problem to be solved is to provide a kind of searching method and device of time domain prediction parameter, can reduce the complexity of interframe encode, improves code efficiency.
In order to solve the problem, the invention discloses a kind of searching method of time domain prediction parameter, comprising:
Carry out merging patterns optimal movement parameter to coding unit to determine, determine the best merging patterns of described coding unit;
Determine that the Adama of the best merging patterns of described coding unit always spends; The Adama of described best merging patterns always spend the Adama cost into described best merging patterns spend with the index of described best merging patterns with, wherein, index cost is the product of the bit number that LaGrange parameter and described best merging patterns index are corresponding;
When the Adama of described best merging patterns always spends satisfied first end condition, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
Preferably, the Adama of described best merging patterns always spends satisfied first end condition, comprising: the Adama of described best merging patterns always spends and is less than first threshold.
Preferably, described first threshold is determined as follows:
Gradient absolute value sum according to current coded unit neighbor calculates the complexity of current coded unit;
Described first threshold is determined according to the complexity of current coded unit and the area of described coding unit and the product of quantization parameter.
Preferably, described method also comprises:
The Adama of described best merging patterns always spend do not meet the first end condition time, calculate total cost that described coding unit is estimated for the one-way movement of first reference picture;
When the Adama of described total cost and described best merging patterns always spends satisfied second end condition, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
Preferably, the Adama of described total cost and described best merging patterns always spends satisfied second end condition to comprise:
Described total cost is greater than Second Threshold; And
The Adama that described total cost is greater than described best merging patterns always spends.
Preferably, described method also comprises:
The Adama of described total cost and described best merging patterns always spend do not meet the second end condition time, perform subsequent motion estimation procedure.
Preferably, described Second Threshold is spend according to the Adama of the complexity of described coding unit, the area of described coding unit and the product of quantization parameter and described best merging patterns the value determined.
According to another aspect of the present invention, provide a kind of searcher of time domain prediction parameter, comprising:
Merging patterns determination module, determines for carrying out merging patterns optimal movement parameter to coding unit, determines the best merging patterns of described coding unit;
First computing module, for determining that the Adama of the best merging patterns of described coding unit always spends; The Adama of described best merging patterns always spend the Adama cost into described best merging patterns spend with the index of described best merging patterns with, wherein, index cost is the product of the bit number that LaGrange parameter and described best merging patterns index are corresponding; And
First determination module, for when the Adama of described best merging patterns always spends satisfied first end condition, determines that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
Preferably, the Adama of described best merging patterns always spends satisfied first end condition, comprising: the Adama of described best merging patterns always spends and is less than first threshold.
Preferably, described device also comprises: for determining the first threshold determination module of first threshold;
Wherein, described first threshold determination module, comprising:
Complicated dynamic behaviour unit, for calculating the complexity of current coded unit according to the gradient absolute value sum of current coded unit neighbor; And
First threshold computing unit, for determining described first threshold according to the complexity of current coded unit and the area of described coding unit and the product of quantization parameter.
Preferably, described device also comprises:
Second computing module, when not meeting the first end condition for always spending at the Adama of described best merging patterns, calculates total cost that described coding unit is estimated for the one-way movement of first reference picture;
Second determination module, for when the Adama of described total cost and described best merging patterns always spends satisfied second end condition, determines that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
Preferably, the Adama of described total cost and described best merging patterns always spends satisfied second end condition to comprise:
Described total cost is greater than Second Threshold; And
The Adama that described total cost is greater than described best merging patterns always spends.
Preferably, described device also comprises:
Motion estimation module, when not meeting the second end condition for always spending at the Adama of described total cost and described best merging patterns, performs subsequent motion estimation procedure.
Preferably, described Second Threshold is spend according to the Adama of the complexity of described coding unit, the area of described coding unit and the product of quantization parameter and described best merging patterns the value determined.
Compared with prior art, the embodiment of the present invention comprises following advantage:
The embodiment of the present invention, when the Adama cost of best merging patterns meets the first end condition, determines that the Prediction Parameters of best merging patterns is the best time domain prediction parameter of current coded unit; Therefore, avoid the motion estimation process of exhaustive merging patterns and all patterns of motion estimation mode and the Adama cost computational process corresponding to pattern, also the amount of calculation of the corresponding motion estimation process of motion estimation mode can namely be saved, reduce computation complexity, thus the complexity of interframe encode can be reduced, improve code efficiency.
Accompanying drawing explanation
Fig. 1 shows the flow chart of steps of the searching method embodiment one of a kind of time domain prediction parameter of the present invention;
Fig. 2 shows the spatial domain prediction schematic diagram under a kind of merging patterns of the present invention;
Fig. 3 shows the time domain prediction schematic diagram under a kind of merging patterns of the present invention;
Fig. 4 shows the flow chart of steps of the searching method embodiment two of a kind of time domain prediction parameter of the present invention;
Fig. 5 shows the flow chart of steps of the searching method embodiment three of a kind of time domain prediction parameter of the present invention;
Fig. 6 shows the flow chart of steps of the searching method example of a kind of time domain prediction parameter of the present invention; And
Fig. 7 shows the structured flowchart of the searcher embodiment of a kind of time domain prediction parameter of the present invention.
Embodiment
For enabling above-mentioned purpose of the present invention, feature and advantage become apparent more, and below in conjunction with the drawings and specific embodiments, the present invention is further detailed explanation.
Present invention improves over the searching method of the time domain prediction parameter of coding unit, current coded unit is being carried out after merging patterns have estimated, obtain the Adama cost of best merging patterns, when the Adama cost of best merging patterns meets the first end condition, can determine that the Prediction Parameters of best merging patterns is the best time domain prediction parameter of current coded unit.Therefore, the motion estimation process and the Adama cost corresponding to all patterns that avoid exhaustive all patterns calculate, and can save the amount of calculation of the corresponding motion estimation process of motion estimation mode, reduce computation complexity, thus the complexity of interframe encode can be reduced, improve code efficiency.
Embodiment one
With reference to Fig. 1, show the flow chart of steps of the searching method embodiment one of a kind of time domain prediction parameter of the present invention, specifically can comprise:
Step 101, coding unit carried out to merging patterns optimal movement parameter and determine, determine the best merging patterns of described coding unit;
In specific implementation, described to coding unit carry out merging patterns optimal movement parameter determine, determine the step of the best merging patterns of described coding unit, specifically can comprise following sub-step:
Sub-step A1, merging patterns action reference variable is carried out to current coded unit, generate 5 merging patterns Prediction Parameters;
In a kind of application example of the present invention, merging patterns action reference variable is carried out to current coded unit, can generate maximum 5 kinds of kinematic parameters (as reference key and motion vector etc.) collection, the concrete number the present invention for motion parameter set is not limited.
Generate merging patterns time domain prediction parameter set and can comprise spatial domain prediction and time domain prediction, if combined prediction and 0 motion vector prediction can be added again not, combined prediction is exactly the Prediction Parameters new according to the temporal motion Prediction Parameters combination producing generated, as the forward prediction parameter of first group and the back forecast parameter of second group form a new prediction, 0 motion vector prediction selects a reference key and 0 set of motion vectors cooperation to be a prediction exactly.With reference to Fig. 2, show the spatial domain prediction schematic diagram under a kind of merging patterns of the present invention, wherein, large rectangle is current block, A0, A1, B0, B1, B2 are encoded reconstructed blocks, and the kinematic parameter of the spatial domain merging patterns of current block is exactly the kinematic parameter of these reference blocks of direct copying, and its order is A1->B1->B0->A0->B2.
With reference to Fig. 3, show the time domain prediction schematic diagram under a kind of merging patterns of the present invention, it finds the block close with current block time-domain position in the reference picture of current block, then using the kinematic parameter of its kinematic parameter as current block.Such as, first, from reference queue, find out the reference frame (being designated as curr-ref) with the POC of present frame (being designated as curr-pic) (display sequence number) immediate frame (being designated as col-pic) and current prediction unit.Secondly, from col-pic, look for two predicting unit alternatively positions, selecting one is time domain reference predicting unit (being designated as col-PU).Finally, that utilize curr-pic with POC distance tb that is curr-ref, the POC distance td of col-pic and col-ref, carries out convergent-divergent to the motion vector of time domain reference predicting unit col-PU, calculates the motion vector of current prediction unit.
Sub-step A2, the Adama cost corresponding to each calculating in described 5 merging patterns Prediction Parameters;
Wherein, the computational process of Adama cost specifically can comprise: for a block of pixels of arbitrary size, if it wide and high is all the multiple of 8, so it is divided into some 8 × 8 sub-blocks, otherwise is divided into some 4 × 4 sub-blocks.Then Hadamard transform is performed to each 8 × 8 sub-blocks or 4 × 4 sub-blocks, again the absolute value of all conversion coefficients is added up and obtain Ap, if current block is 4 × 4, so (Ap+1)/2 are exactly the Adama cost of 4 × 4 pieces, if 8 × 8 pieces, so (Ap+2)/4 be exactly current 8 × 8 pieces Adama cost, a block Adama cost be its all 4 × 4 or 8 × 8 sub-blocks be partitioned into Adama cost and.
Sub-step A3, determine that minimum Adama always spends corresponding merging patterns to be the best merging patterns of current coded unit.
Particularly, best merging (Merge) pattern can by minimizing HAD
m+ bit
mrgidx× λ
motiondecide, wherein, HAD
mfor the Adama cost that best merging patterns are corresponding, bit
mrgidxfor the code check of best merging patterns index spends, for 5 Merge indexes, their code check cost is not identical, λ
motionfor LaGrange parameter.In embodiments of the present invention, by bit
mrgidx× λ
motionbe designated as the index cost that best merging patterns are corresponding, namely the Adama of best merging patterns always spends the Adama cost HAD being specifically as follows described best merging patterns
mbit is spent with corresponding index
mrgidx× λ
motionand.
Step 102, determine that the Adama of the best merging patterns of described coding unit always spends; The Adama of described best merging patterns always spend the Adama cost into described best merging patterns spend with the index of described best merging patterns with, wherein, index cost is the product of the bit number that LaGrange parameter and described best merging patterns index are corresponding;
According to the result of calculation in abovementioned steps 101, can determine that minimum Adama always spends the Adama for the best merging patterns of current coded unit always to spend.
Step 103, when the Adama of described best merging patterns always spends satisfied first end condition, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
In the embodiment of the present invention, first end condition may be used for representing that the Adama of described best merging patterns always spends enough little various conditions, when the Adama of described best merging patterns always spends satisfied first end condition, can think that the cost of described best merging patterns can be born, therefore without the need to carrying out subsequent motion estimation procedure corresponding to motion estimation mode again.
Particularly, the Adama of described best merging patterns always spends satisfied first end condition, specifically can comprise: the Adama of described best merging patterns always spends and is less than first threshold.
Such as, can determine that the Adama of best merging patterns always spends by such as lower inequality and whether meet the first end condition:
th
fst＞HAD
m+bit
mrgidx×λ
motion(1)
Namely when the Adama of described best merging patterns always spend meet above-mentioned inequality (1) time, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit, follow-up motion estimation process can be stopped.Wherein, first threshold th
fstit can be a less value, therefore above-mentioned inequality illustrates that the cost of best merging patterns is very little, so can think that best merging patterns are best interframe encoding mode, also namely the Prediction Parameters of best merging patterns is the best time domain prediction parameter of current coded unit, therefore without the need to carrying out the estimation of other pattern again.Wherein, first threshold th
fstcan be obtained by encoder experiment statistics, λ
motionfor LaGrange parameter, bit
mrgidxfor the bit number that merging patterns index needs.
When meeting the first end condition, follow-up motion estimation process can be stopped, therefore can get rid of the amount of calculation of the estimation of a part, thus reducing computation complexity.
In a kind of embodiment of the present invention, first threshold th
fstcan be determined by following sub-step:
The gradient absolute value sum of sub-step B1, foundation current coded unit neighbor calculates the complexity of current coded unit;
Specifically can by the complexity grad of following formulae discovery current coded unit:
Wherein, I
i,jdenotation coordination is (i, j) place image pixel intensities, and abs is absolute value operation, and Ω is the set of all coordinate figures of allowing in current coded unit.
Sub-step B2, product determination first threshold th according to the complexity of current coded unit and the area of described coding unit and quantization parameter
fst;
th
fst＝(((grad×5＞＞2)+(M×N×qp×9＞＞7))×3)＞＞4 (3)
By the first threshold th that above formula (3) calculates
fstcan be a less value, wherein, qp be the quantization parameter of current coded unit, and grad × 5 > > 2 represents that the gradient of neighbor is larger, first threshold th
fstalso larger.M × N × qp is the product of area and quantization parameter, represents that quantization parameter is larger, first threshold th
fstlarger.
It should be noted that, the Adama of best merging patterns described above always spends and is less than first threshold just as an example of the first end condition, those skilled in the art can adopt other the first end condition according to actual needs, and the embodiment of the present invention is not limited the first concrete end condition.
To sum up, the embodiment of the present invention is being carried out after merging patterns action reference variable completes to current coded unit, the Adama obtaining best merging patterns corresponding always spends, when the Adama that these best merging patterns are corresponding always spends satisfied first end condition, namely can determine that the Prediction Parameters of these best merging patterns is the best time domain prediction parameter of current coded unit.Therefore, the motion estimation process and the Adama cost corresponding to all patterns that avoid exhaustive all patterns calculate, and can save the amount of calculation of the corresponding motion estimation process of motion estimation mode, reduce computation complexity, thus the complexity of interframe encode can be reduced, improve code efficiency.
Embodiment two
With reference to Fig. 4, show the flow chart of steps of the searching method embodiment two of a kind of time domain prediction parameter of the present invention, specifically can comprise:
Step 401, coding unit carried out to merging patterns optimal movement parameter and determine, determine the best merging patterns of described coding unit;
Step 402, determine that the Adama of the best merging patterns of described coding unit always spends; The Adama of described best merging patterns always spend the Adama cost into described best merging patterns spend with the index of described best merging patterns with, wherein, index cost is the product of the bit number that LaGrange parameter and described best merging patterns index are corresponding;
Step 403, when the Adama of described best merging patterns always spends satisfied first end condition, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit;
The searching method of the time domain prediction parameter of the present embodiment, on the basis of above-described embodiment one, can also comprise the steps: further
Step 404, the Adama of described best merging patterns always spend do not meet the first end condition time, calculate total cost that described coding unit is estimated for the one-way movement of first reference picture;
Step 405, when the Adama of described total cost and described best merging patterns always spends satisfied second end condition, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
The Adama of described best merging patterns always spend do not meet the first end condition time, need to continue to perform the best time domain prediction parameter that follow-up one-way movement estimates to determine current coded unit.
Wherein, described one-way movement estimation procedure specifically can comprise:
Current coded unit is mated in first reference picture, and find that block mated most with current coded unit, this block may point to fractional position, in this case, can calculate described one-way movement estimate corresponding prediction block by interpolation filter.
Calculate total cost that described coding unit is estimated for the one-way movement of first reference picture, specifically can comprise following sub-step:
Sub-step C1, estimation is carried out to first reference picture, obtain motion vector, determine that corresponding code check cost bits is estimated in described one-way movement
mv0(specifically can comprise motion vector and reference key);
Wherein, motion vector is side-play amount, and reference key is for identifying concrete which secondary reference picture of sensing.
Sub-step C2, determine described unidirectional estimation matching error Adama cost HAD
mv0;
Wherein, matching error is that unit to be encoded and corresponding predicting unit subtract each other the difference obtained, and then carries out Hadamard transform to this difference, obtains the Adama cost HAD of matching error
mv0.
Sub-step C3, according to described code check cost bits
mv0with the Adama cost HAD of matching error
mv0calculate the described unidirectional total cost cost estimating correspondence
mv0.
Particularly, always cost is spent
mv0can be calculated by following formula:
cost
mv0＝HAD
mv0+λ
motion×bits
mv0(4)
At described total cost cost
mv0when always spending satisfied second end condition with the Adama of described best merging patterns, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.Meanwhile, follow-up motion estimation process can be stopped.
In a kind of embodiment of the present invention, the Adama of described total cost and described best merging patterns always spends satisfied second end condition to comprise:
Described total cost is greater than Second Threshold; And
The Adama that described total cost is greater than described best merging patterns always spends.
Particularly, when following two inequality are set up simultaneously, can determine that the Adama of described total cost and described best merging patterns always spends satisfied second end condition:
cost
mv0＞th
scd(5)
cost
mv0＞HAD
m+bit
mrgidx×λ
motion(6)
Wherein, cost
mv0for total cost that current coded unit is estimated for the one-way movement of first reference picture, th
scdfor Second Threshold, HADm is the Adama cost of the best merging patterns of current coded unit, λ
motionfor LaGrange parameter, bit
mrgidxfor the bit number that merging patterns index needs, HAD
m+ bit
mrgidx× λ
motionfor the Adama of the best merging patterns of current coded unit always spends.If above-mentioned two inequality (5) and (6) are set up simultaneously, namely the second end condition is met, then determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit, and subsequent motion estimation procedure can be stopped.
When meeting the second end condition, follow-up motion estimation process can be stopped, the amount of calculation of the estimation of a part can be got rid of, thus reduce computation complexity.
Wherein, Second Threshold can for spending according to the complexity of described coding unit, the area of described coding unit and the product of quantization parameter and described Adama the value determined.
Particularly, following formulae discovery Second Threshold th can be passed through
scd:
th
scd＝(((grad×53＞＞7)-(29×M×N×qp＞＞9)+(37×HAD
m＞＞5))×19)＞＞4 (7)
Wherein, grad × 53 > > 7 represents that the gradient of neighbor is larger, Second Threshold th
scdalso larger; 29 × M × N × qp > > 9 represents that quantization parameter is larger, and so encoder more tends to select Merge pattern, so quantization parameter is larger, and th
scdless; 37 × HAD
m> > 5 does weighting to the HADm item of best Merge pattern, represents that the HADm of best Merge pattern is larger, so th
scdlarger.
Second Threshold th
scdmay be used for the judgement to subsequent motion estimation procedure, can represent if first reference picture of current coded unit at least should draw the cost less than current cost, just have the meaning continuing to perform subsequent motion estimation.Namely when meeting the second end condition, the cost cost of the first reference picture
mv0the Adama being greater than best Merge pattern always spends, and cost
mv0than a threshold value th
scdalso large, this threshold value th
scdwith the linear relation of best Merge pattern, see above-mentioned th
scdcomputational methods, th
scdthe Adama of best Merge pattern cost HADm is also exaggerated, that is, the cost of the first reference picture does not only reduce cost total value, also more much bigger than the cost of best Merge pattern on the contrary, so do not think that other reference pictures can reduce cost yet, therefore, follow-up motion estimation process can be stopped, determine that the Prediction Parameters of best Merge pattern is the best time domain prediction parameter of current coded unit.Above-mentioned formula can be specifically the ratio of the optimum that the data statistics produced by encoder is drawn, object reduces computing time as much as possible.
To sum up, the present embodiment the Adama of best merging patterns always spend not meet the first end condition time, if the Adama of total cost that current coded unit is estimated for the one-way movement of first reference picture and best merging patterns always spends satisfied second end condition, then determine that the Prediction Parameters of these best merging patterns is the best time domain prediction parameter of current coded unit.Follow-up more unidirectional and bi-directional motion estimation process can be avoided, therefore can reduce computation complexity further, thus the complexity of pre-interframe encode can be reduced, improve code efficiency.
Embodiment three
With reference to Fig. 5, the flow chart of steps showing the searching method embodiment three of a kind of time domain prediction parameter of the present invention, specifically can comprise:
Step 501, coding unit carried out to merging patterns optimal movement parameter and determine, determine the best merging patterns of described coding unit;
Step 502, determine that the Adama of the best merging patterns of described coding unit always spends; The Adama of described best merging patterns always spend the Adama cost into described best merging patterns spend with the index of described best merging patterns with, wherein, index cost is the product of the bit number that LaGrange parameter and described best merging patterns index are corresponding;
Step 503, when the Adama of described best merging patterns always spends satisfied first end condition, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit;
Step 504, the Adama of described best merging patterns always spend do not meet the first end condition time, calculate total cost that described coding unit is estimated for the one-way movement of first reference picture;
Step 505, when the Adama of described total cost and described best merging patterns always spends satisfied second end condition, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit;
The searching method of the time domain prediction parameter of the present embodiment, on the basis of above-described embodiment two, can also comprise the steps: further
Step 506, the Adama of described total cost and described best merging patterns always spend do not meet the second end condition time, perform subsequent motion estimation procedure.
Particularly, the Adama of described total cost and described best merging patterns always spend do not meet the second end condition time, continue to perform subsequent motion estimation procedure, obtain best forward and backward kinematic parameter, then perform bi-directional motion estimation and obtain best bidirectional-movement parameter, from best forward, backward and bidirectional-movement parameter three, select optimum one to calculate Adama always spend, the Adama of result of calculation and best merging patterns is always spent and contrasts, determine that the Prediction Parameters of the associative mode of the total least cost of Adama is best time domain prediction parameter.
Embodiment four
With reference to Fig. 6, show the flow chart of steps of the searching method example of a kind of time domain prediction parameter of the present invention, specifically can comprise:
Step 601, current coded unit carried out to Merge pattern optimal movement parameter and determine, determine the best Merge pattern of current coded unit;
Step 602, determine that the Adama of the best Merge pattern of current coded unit always spends; The Adama of described best merging patterns always spend the Adama cost into described best merging patterns spend with the index of described best merging patterns with, wherein, index cost is the product of the bit number that LaGrange parameter and described best merging patterns index are corresponding;
Step 603, when the Adama of best Merge pattern always spends satisfied first end condition, determine that the Prediction Parameters of best Merge pattern is the best time domain prediction parameter of current coded unit;
Step 604, the Adama of best Merge pattern always spend do not meet the first end condition time, calculate total cost cost that current coded unit is estimated for the one-way movement of first reference picture
mv0;
Step 605, at cost
mv0when always spending satisfied second end condition with the Adama of described best Merge pattern, determine that the Prediction Parameters of described best Merge pattern is the best time domain prediction parameter of current coded unit;
Step 606, at cost
mv0with the Adama of described best Merge pattern always spend do not meet the second end condition time, perform subsequent motion estimation procedure.
Wherein, the first end condition, the second end condition, Adama always spend and cost
mv0computational process see embodiment one, no longer can repeat herein.
Device embodiment
With reference to Fig. 7, show the structured flowchart of the searcher embodiment of a kind of time domain prediction parameter of the present invention, specifically can comprise:
Merging patterns determination module 710, determines for carrying out merging patterns optimal movement parameter to coding unit, determines the best merging patterns of described coding unit;
First computing module 720, for determining that the Adama of the best merging patterns of described coding unit always spends; The Adama of described best merging patterns always spend the Adama cost into described best merging patterns spend with the index of described best merging patterns with, wherein, index cost is the product of the bit number that LaGrange parameter and described best merging patterns index are corresponding; And
First determination module 730, for when the Adama of described best merging patterns always spends satisfied first end condition, determines that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
In one preferred embodiment of the invention, the Adama of described best merging patterns always spends satisfied first end condition, can comprise: the Adama of described best merging patterns always spends and is less than first threshold.
In another preferred embodiment of the invention, described device can also comprise: for determining the first threshold determination module of first threshold;
Wherein, described first threshold determination module, can comprise:
Complicated dynamic behaviour unit, for calculating the complexity of current coded unit according to the gradient absolute value sum of current coded unit neighbor; And
First threshold computing unit, for determining described first threshold according to the complexity of current coded unit and the area of described coding unit and the product of quantization parameter.
In another preferred embodiment of the present invention, described device can also comprise:
Second computing module, when not meeting the first end condition for always spending at the Adama of described best merging patterns, calculates total cost that described coding unit is estimated for the one-way movement of first reference picture;
Second determination module, for when the Adama of described total cost and described best merging patterns always spends satisfied second end condition, determines that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
In another preferred embodiment of the present invention, the Adama of described total cost and described best merging patterns always spends satisfied second end condition to comprise:
Described total cost is greater than Second Threshold; And
The Adama that described total cost is greater than described best merging patterns always spends.
In another preferred embodiment of the present invention, described device can also comprise:
Motion estimation module, when not meeting the second end condition for always spending at the Adama of described total cost and described best merging patterns, can perform subsequent motion estimation procedure.
In another preferred embodiment of the present invention, described Second Threshold can for spending according to the Adama of the complexity of described coding unit, the area of described coding unit and the product of quantization parameter and described best merging patterns the value determined.
For device embodiment, due to itself and embodiment of the method basic simlarity, so description is fairly simple, relevant part illustrates see the part of embodiment of the method.
Each embodiment in this specification all adopts the mode of going forward one by one to describe, and what each embodiment stressed is the difference with other embodiments, between each embodiment identical similar part mutually see.
Those skilled in the art should understand, the embodiment of the embodiment of the present invention can be provided as method, device or computer program.Therefore, the embodiment of the present invention can adopt the form of complete hardware embodiment, completely software implementation or the embodiment in conjunction with software and hardware aspect.And the embodiment of the present invention can adopt in one or more form wherein including the upper computer program implemented of computer-usable storage medium (including but not limited to magnetic disc store, CD-ROM, optical memory etc.) of computer usable program code.
The embodiment of the present invention describes with reference to according to the flow chart of the method for the embodiment of the present invention, terminal equipment (system) and computer program and/or block diagram.Should understand can by the combination of the flow process in each flow process in computer program instructions realization flow figure and/or block diagram and/or square frame and flow chart and/or block diagram and/or square frame.These computer program instructions can being provided to the processor of all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing terminal equipment to produce a machine, making the instruction performed by the processor of computer or other programmable data processing terminal equipment produce device for realizing the function of specifying in flow chart flow process or multiple flow process and/or block diagram square frame or multiple square frame.
These computer program instructions also can be stored in can in the computer-readable memory that works in a specific way of vectoring computer or other programmable data processing terminal equipment, the instruction making to be stored in this computer-readable memory produces the manufacture comprising command device, and this command device realizes the function of specifying in flow chart flow process or multiple flow process and/or block diagram square frame or multiple square frame.
These computer program instructions also can be loaded on computer or other programmable data processing terminal equipment, make to perform sequence of operations step to produce computer implemented process on computer or other programmable terminal equipment, thus the instruction performed on computer or other programmable terminal equipment is provided for the step realizing the function of specifying in flow chart flow process or multiple flow process and/or block diagram square frame or multiple square frame.
Although described the preferred embodiment of the embodiment of the present invention, those skilled in the art once obtain the basic creative concept of cicada, then can make other change and amendment to these embodiments.So claims are intended to be interpreted as comprising preferred embodiment and falling into all changes and the amendment of embodiment of the present invention scope.
Finally, also it should be noted that, in this article, the such as relational terms of first and second grades and so on is only used for an entity or operation to separate with another entity or operating space, and not necessarily requires or imply the relation that there is any this reality between these entities or operation or sequentially.And, term " comprises ", " comprising " or its any other variant are intended to contain comprising of nonexcludability, thus make to comprise the process of a series of key element, method, article or terminal equipment and not only comprise those key elements, but also comprise other key elements clearly do not listed, or also comprise by the intrinsic key element of this process, method, article or terminal equipment.When not more restrictions, the key element limited by statement " comprising ... ", and be not precluded within process, method, article or the terminal equipment comprising described key element and also there is other identical element.
Above to searching method and the device of a kind of time domain prediction parameter provided by the present invention, be described in detail, apply specific case herein to set forth principle of the present invention and execution mode, the explanation of above embodiment just understands method of the present invention and core concept thereof for helping; Meanwhile, for one of ordinary skill in the art, according to thought of the present invention, all will change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.
Claims (14)
1. a searching method for time domain prediction parameter, is characterized in that, described method comprises:
Carry out merging patterns optimal movement parameter to coding unit to determine, determine the best merging patterns of described coding unit;
Determine that the Adama of the best merging patterns of described coding unit always spends; The Adama of described best merging patterns always spend the Adama cost into described best merging patterns spend with the index of described best merging patterns with, wherein, index cost is the product of the bit number that LaGrange parameter and described best merging patterns index are corresponding;
When the Adama of described best merging patterns always spends satisfied first end condition, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
2. the method for claim 1, is characterized in that, the Adama of described best merging patterns always spends satisfied first end condition, comprising: the Adama of described best merging patterns always spends and is less than first threshold.
3. method as claimed in claim 2, is characterized in that, determine described first threshold as follows:
Gradient absolute value sum according to current coded unit neighbor calculates the complexity of current coded unit;
Described first threshold is determined according to the complexity of current coded unit and the area of described coding unit and the product of quantization parameter.
4. the method for claim 1, is characterized in that, described method also comprises:
The Adama of described best merging patterns always spend do not meet the first end condition time, calculate total cost that described coding unit is estimated for the one-way movement of first reference picture;
When the Adama of described total cost and described best merging patterns always spends satisfied second end condition, determine that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
5. method as claimed in claim 4, it is characterized in that, the Adama of described total cost and described best merging patterns always spends satisfied second end condition to comprise:
Described total cost is greater than Second Threshold; And
The Adama that described total cost is greater than described best merging patterns always spends.
6. the method as described in claim 4 or 5, is characterized in that, described method also comprises:
The Adama of described total cost and described best merging patterns always spend do not meet the second end condition time, perform subsequent motion estimation procedure.
7. the method as described in claim 4 or 5, is characterized in that, described Second Threshold is spend according to the Adama of the complexity of described coding unit, the area of described coding unit and the product of quantization parameter and described best merging patterns the value determined.
8. a searcher for time domain prediction parameter, is characterized in that, described device comprises:
Merging patterns determination module, determines for carrying out merging patterns optimal movement parameter to coding unit, determines the best merging patterns of described coding unit;
First computing module, for determining that the Adama of the best merging patterns of described coding unit always spends; The Adama of described best merging patterns always spend the Adama cost into described best merging patterns spend with the index of described best merging patterns with, wherein, index cost is the product of the bit number that LaGrange parameter and described best merging patterns index are corresponding; And
First determination module, for when the Adama of described best merging patterns always spends satisfied first end condition, determines that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
9. device as claimed in claim 8, it is characterized in that, the Adama of described best merging patterns always spends satisfied first end condition, comprising: the Adama of described best merging patterns always spends and is less than first threshold.
10. device as claimed in claim 9, it is characterized in that, described device also comprises: for determining the first threshold determination module of first threshold;
Wherein, described first threshold determination module, comprising:
Complicated dynamic behaviour unit, for calculating the complexity of current coded unit according to the gradient absolute value sum of current coded unit neighbor; And
First threshold computing unit, for determining described first threshold according to the complexity of current coded unit and the area of described coding unit and the product of quantization parameter.
11. devices as claimed in claim 8, it is characterized in that, described device also comprises:
Second computing module, when not meeting the first end condition for always spending at the Adama of described best merging patterns, calculates total cost that described coding unit is estimated for the one-way movement of first reference picture;
Second determination module, for when the Adama of described total cost and described best merging patterns always spends satisfied second end condition, determines that the Prediction Parameters of described best merging patterns is the best time domain prediction parameter of current coded unit.
12. devices as claimed in claim 11, is characterized in that, the Adama of described total cost and described best merging patterns always spends satisfied second end condition to comprise:
Described total cost is greater than Second Threshold; And
The Adama that described total cost is greater than described best merging patterns always spends.
13. devices as described in claim 11 or 12, it is characterized in that, described device also comprises:
Motion estimation module, when not meeting the second end condition for always spending at the Adama of described total cost and described best merging patterns, performs subsequent motion estimation procedure.
14. devices as described in claim 11 or 12, is characterized in that, described Second Threshold is spend according to the Adama of the complexity of described coding unit, the area of described coding unit and the product of quantization parameter and described best merging patterns the value determined.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410851138.2A CN104601991B (en) | 2014-12-30 | 2014-12-30 | A kind of searching method and device of time domain prediction parameter |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410851138.2A CN104601991B (en) | 2014-12-30 | 2014-12-30 | A kind of searching method and device of time domain prediction parameter |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104601991A true CN104601991A (en) | 2015-05-06 |
| CN104601991B CN104601991B (en) | 2018-12-18 |
Family
ID=53127446
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410851138.2A Active CN104601991B (en) | 2014-12-30 | 2014-12-30 | A kind of searching method and device of time domain prediction parameter |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104601991B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104918047A (en) * | 2015-05-29 | 2015-09-16 | 北京奇艺世纪科技有限公司 | Bidirectional motion estimation elimination method and device |
| CN107135396A (en) * | 2017-06-16 | 2017-09-05 | 北京奇艺世纪科技有限公司 | A kind of method for estimating and device |
| CN108156458A (en) * | 2017-12-28 | 2018-06-12 | 北京奇艺世纪科技有限公司 | A kind of determining coding mode method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014106327A1 (en) * | 2013-01-04 | 2014-07-10 | Mediatek Singapore Pte. Ltd. | Method and apparatus for inter-view residual prediction in multiview video coding |
| CN104023234A (en) * | 2014-06-24 | 2014-09-03 | 华侨大学 | Fast inter-frame prediction method applicable to high efficiency video coding (HEVC) |
| CN104113754A (en) * | 2014-07-17 | 2014-10-22 | 四川大学 | Method for high-performance video interframe coding based on time domain relevance and transcoder thereof |
-
2014
- 2014-12-30 CN CN201410851138.2A patent/CN104601991B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014106327A1 (en) * | 2013-01-04 | 2014-07-10 | Mediatek Singapore Pte. Ltd. | Method and apparatus for inter-view residual prediction in multiview video coding |
| CN104023234A (en) * | 2014-06-24 | 2014-09-03 | 华侨大学 | Fast inter-frame prediction method applicable to high efficiency video coding (HEVC) |
| CN104113754A (en) * | 2014-07-17 | 2014-10-22 | 四川大学 | Method for high-performance video interframe coding based on time domain relevance and transcoder thereof |
Non-Patent Citations (4)
| Title |
|---|
| LIQUAN SHEN,ZHAOYANG ZHANG,ZHI LIU: "《Adaptive Inter-Mode Decision for HEVC Jointly Utilizing Inter-Level and Spatiotemporal Correlations》", 《IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHN0LOGY》 * |
| MIOK KIM,HYUK-JAE LEE,NAM LING: "《fast mode decision for diamond search in high efficiency video coding》", 《VISUAL COMMUNICATIONS AND IMAGE PROCESSING(VICP),2013》 * |
| ZHAOQING PAN,SAM KWONG,MING-TING SUN,JIANJUN LEI: "《Early MERGE Mode Decision Based on Motion Estimation and Hierarchical Depth Correlation for HEVC》", 《IEEE TRANSACTIONS ON BROADCASTING》 * |
| 罗霁: "《HEVC快速CU大小选择与SKIP模式早期判断算法》", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104918047A (en) * | 2015-05-29 | 2015-09-16 | 北京奇艺世纪科技有限公司 | Bidirectional motion estimation elimination method and device |
| CN107135396A (en) * | 2017-06-16 | 2017-09-05 | 北京奇艺世纪科技有限公司 | A kind of method for estimating and device |
| CN107135396B (en) * | 2017-06-16 | 2020-08-28 | 北京奇艺世纪科技有限公司 | Motion estimation method and device |
| CN108156458A (en) * | 2017-12-28 | 2018-06-12 | 北京奇艺世纪科技有限公司 | A kind of determining coding mode method and device |
| CN108156458B (en) * | 2017-12-28 | 2020-04-10 | 北京奇艺世纪科技有限公司 | Method and device for determining coding mode |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104601991B (en) | 2018-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10057597B2 (en) | Method and apparatus for encoding and decoding motion vector based on reduced motion vector predictor candidates | |
| KR102073638B1 (en) | Picture prediction method and picture prediction device | |
| KR102013771B1 (en) | Image prediction method and related device | |
| CN102835111B (en) | The motion vector of previous block is used as the motion vector of current block, image to be carried out to the method and apparatus of coding/decoding | |
| CN106686378B (en) | The method and apparatus that image is decoded | |
| CN102598670A (en) | Method and apparatus for encoding/decoding image with reference to a plurality of frames | |
| CN100473168C (en) | Motion vector space prediction method for video coding | |
| CN106105191A (en) | For the method and apparatus processing multiview video signal | |
| MX2012011646A (en) | Method and apparatus for performing interpolation based on transform and inverse transform. | |
| CN104918053A (en) | Methods and apparatuses for encoding and decoding motion vector | |
| CN104469362A (en) | Method and apparatus for encoding and decoding motion vector | |
| CN104754338A (en) | Selection method and device for intra-frame predication mode | |
| KR20180037042A (en) | A motion vector field coding method and a decoding method, and a coding and decoding apparatus | |
| CN104601991A (en) | Searching method and device for time-domain predictive parameters | |
| CN104918047B (en) | A kind of method and device for removing of bi-directional motion estimation | |
| CN105025298A (en) | A method and device of encoding/decoding an image | |
| CN103491371A (en) | Encoding method, device and equipment based on hierarchy | |
| CN110475116A (en) | A kind of motion vector deriving method, device and electronic equipment | |
| CN105611299A (en) | Motion estimation method based on HEVC | |
| CN103002279B (en) | Encode the method and its corresponding device of simultaneously reconstructed pixel block | |
| JP6935553B2 (en) | Methods and Devices for Coding / Decoding Motion Vectors Based on Reduced Predicted Motion Vector Candidates | |
| JP6872412B2 (en) | Video coding device and program | |
| CN106254884A (en) | Inter-frame prediction method based on H265 coding and encoder IP | |
| CN105025299A (en) | Method and apparatus for encoding/decoding image | |
| CN103581689A (en) | Video coding method and video coding device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |